The DNA replication machinery transmits dual signals to prevent unscheduled licensing and execution of centrosome duplication

Abstract

Copy number control of DNA and centrosomes is essential for accurate genetic inheritance. DNA replication and centrosome duplication have been recognized as parallel key events for cell division. Here, we discover that the DNA replication machinery directly regulates the licensing and execution processes of centrosome duplication to prevent centrosome amplification. We find that the microcephaly protein DONSON couples DNA replication initiation with Cdc6 translocation to centrosomes. The Cdc6 signal prevents the precocious occurrence of centriole disengagement, the licensing step for centrosome duplication. During DNA replication, DONSON inhibits replisome disassembly by interacting with the CMG helicase, maintaining the intrinsic S/G2 checkpoint signal that blocks centriole-to-centrosome conversion, the execution step for centrosome duplication. Disruption of these dual signals causes precocious centrosome duplication and chromosome mis-segregation, observed in DONSON patient cells. Our results reveal that the DNA replication machinery not only duplicates genetic material but also controls the system for its accurate segregation.

Fig. 1. DONSON and Cdc6, DNA replication factors, maintain centriole engagement and proper centrosome number.

a List of genes involved in microcephalic primordial dwarfism (MPD) based on previous studies. The genes involved in MPD are mostly classified as centrosomal proteins, DNA replication components, or a regulator for DNA damage response. b The number of Cep192 foci upon siRNA treatment against the indicated proteins. c, Histograms represent the frequency of interphase cells with >2 Cep192 foci observed in (b). n = 3 independent experiments, 50 cells for each. d Schematic of the DONSON gene, indicating mutations observed in cells from MPD patient 2 (P2). The genomic structure is based on the longest ORF, containing ten coding exons (white rectangles) (NCBI NM_017613.3). e The number of Cep192 foci in hTERT-immortalized fibroblasts derived from the MPD patient 2. Fibroblasts were infected with pMSCV-empty-vector or pMSCV-DONSON and immunostained with antibodies against Cep192 (red). Arrowheads indicate centrosomes. f Histograms represent the frequency of interphase cells with the indicated phenotypes observed in (e). n = 3 independent experiments, 50 cells for each. g The number of centrin and Cep192 foci in G1 or S/G2 phase upon DONSON or Cdc6 depletion. HeLa cells were treated with siRNA treatment against the indicated proteins for 48 h and 10 μM EdU for 30 min. CENP-F and EdU double-negative cells were classified as G1 phase cells, and others as S or G2 phase cells. Arrowheads indicate centrioles. h Histograms represent the frequency of cells with the indicated phenotype observed in (g). n = 3 independent experiments, 30 cells for each. i Time-lapse observation of centriole disengagement in DONSON-depleted cells. HeLa GFP-centrin-1 cells were visualized every 20 min for 48 h after 24 h treatment with siRNA and 3 h treatment with 100 nM SiR-DNA. Arrowheads indicate engaged centriole pairs, and left–right arrows indicate precociously disengaged centrioles. j Cumulative scatterplot indicates the duration from nuclear envelope breakdown (NEBD) to centriole disengagement observed in (i). k Schematic model of the phenotypes of DONSON or Cdc6 depletion. All scale bars, 5 μm. Values are mean percentages ± s.d. Tukey’s multiple comparison test was used in c, and a two-tailed, unpaired Student’s t-test was used in (f) to obtain P-value. Throughout all the figures in this study, the single-channel images were generated by displaying only the signal from one channel extracted directly from the original merged representative image, without any adjustment to intensity or contrast. The insets represent magnified views of regions taken directly from either the merged image or the corresponding single-channel image, also without any further modification. The insets are displayed in the following order from top to bottom: merged, green, red, and cyan. All source data are provided as a Source Data file.

Fig. 2. DONSON maintains centriole engagement through the translocation of Cdc6 from the nucleus to the centrosome.

a Subcellular localization of overexpressed DONSON and Cdc6 in G1 and S/G2 phases. HeLa-Tet3G cells were treated with 1 μg/mL Doxycycline (Dox) for 24 h. b HeLa cells were fixed 24 h after synchronization in G1 phase using 10 µM PHA767491 (Cdc7i) or 4 h after release. c Cdc6 localization at centrioles in S phase in control or DONSON-depleted cells. d Quantification of the normalized signal intensity of Cdc6 at each centriole observed in (c). n  =  50 from three independent experiments. e Inducible expression of Cdc6 upon DONOSN depletion. HeLa-Tet3G-mNG-Cdc6 cells were treated with siRNA for 48 h, followed by 1 μg/mL Dox treatment for 24 h. f Histograms represent the frequency of interphase cells with > 2 C-Nap1 foci observed in (e). n  =  3 independent experiments, 30 cells for each. g List of proteins previously reported to be involved in centriole engagement or disengagement. h Precocious centriole disengagement was inhibited upon co-depletion of Cdh1 and DONSON/Cdc6. i, j Histograms represent the frequency of interphase cells with >2 C-Nap1 foci observed in (h). n  =  3 independent experiments, 50 cells for each. k Cdh1 localization at centrosomes in cells treated with siControl or siCdc6. l Inducible expression of Cdc6 upon DONOSN depletion. HeLa-Tet3G-mRuby3-Cdc6 cells were treated with siRNA for 48 h, followed by 1 μg/mL Dox treatment for 24 h. m Quantification of the normalized signal intensity of Cdh1 at each centriole with the indicated treatment. n = 50 from three independent experiments. n Schematic illustration of the phenotype of DONSON depletion. The depletion of DONSON leads to the mislocalization of Cdc6 at the centrosomes during the S phase, where Cdc6 inhibits the premature recruitment of Cdh1. All scale bars, 5 μm. Values are mean percentages ± s.d. Tukey’s multiple comparison test was used in (i), Two-tailed, unpaired Student’s t-test was used in (f) and (j), Mann–Whitney U test was used in (d), and Kruskal–Wallis test was used in (m) to obtain P-value. Source data are provided as a Source Data file.

Fig. 3. DONSON suppresses precocious Plk1 activation and prevents unscheduled maturation of daughter centrioles during the S phase.

a Recruitment of PCNT to disengaged daughter centrioles in S phase after siRNA treatment against the indicated proteins. After 32 h of siRNA transfection, cells were treated with EdU for 30 min to classify cells in S phase. b Histograms represent the frequency of cells with more than two PCNT-positive centrioles in S phase observed in (a). n  =  3 independent experiments, 30 cells for each. c Ectopic MTOC activity of precociously disengaged daughter centrioles in S phase in DONSON-depleted cells. HeLa cells were treated with siControl, siCep57/Cep57L1, or siDONSON for 32 h and followed by 10 µM nocodazole treatment for 3 h. After nocodazole treatment, the cells were cold-treated for 1 h, followed by 30-s incubation at 37 °C. d Phosphorylation of Plk1 at T210 residue at centrosomes of WT cells in G1, S, and G2 phases. e Phosphorylation of Plk1 at T210 residue in S phase upon DONSON depletion. f Quantification of the normalized signal intensity of phosphorylated Plk1 (T210) at each centriole observed in (e). n = 50 from three independent experiments. g Recruitment of γ-Tubulin to disengaged daughter centrioles upon Plk1 inhibition in DONSON-depleted cells. HeLa cells were treated with siControl or siDONSON for 48 h, followed by treatment with DMSO or 100 nM BI2536 (Plk1i) for 24 h. After siDONSON treatment, Plk1 activity remains high, preventing linker formation between disengaged centrioles. In contrast, DONSON-depleted cells with Plk1 inhibition retain the linker after disengagement, keeping centrioles closely spaced. Disengagement is defined as >2 C-Nap1 foci. h, i Histograms represent the frequency of the interphase cells with the indicated phenotypes observed in (g). n  =  3 independent experiments, 50 cells for each. j The basal Chk1 activity upon DONSON depletion. Representative immunoblot analysis of whole cell extracts for Chk1 and p-Chk1 (S296) from HeLa cells treated with siControl or siDONSON for 48 h. HSP90 was used as a loading control. k Overexpression of a constitutively active mutant of Chk1 (Chk1-L449R construct) upon DONSON depletion. l Histograms represent the frequency of cells with more than two γ-Tubulin-positive centrioles observed in (k). n  =  3 independent experiments, 30 cells for each. m Schematic model of the phenotypes of DONSON depletion. All scale bars, 5 μm. Values are mean percentages ± s.d. Tukey’s multiple comparison test was used in (h) and (i), a two-tailed, unpaired Student’s t-test was used in (l), and Mann–Whitney U test was used in (f) to obtain the P-value. Source data are provided as a Source Data file.

Fig. 4. DONSON regulates the intrinsic S/G2 checkpoint by suppressing unscheduled ETAA1 degradation.

a A schematic of the intrinsic S/G2 checkpoint based on previous studies^,,. b The subcellular localization pattern of CyclinB1 of WT cells in G1, S, and G2 phases. c The expression pattern of CyclinB1 in S phase upon Chk1- or ATR-inhibition. HeLa cells were treated with 250 nM CHIR-124 (Chk1i) or 2 μM VE-821 (ATRi) for 6 h and EdU for 30 min to classify cells in S phase. d Histograms represent the frequency of the S phase cells with the indicated phenotypes observed in (c). n  =  3 independent experiments, 30 cells for each. e The expression pattern of CyclinB1 in S phase after siRNA treatment against the indicated proteins. f Histograms represent the frequency of the S phase cells with the indicated phenotypes observed in (e). n  =  3 independent experiments, 30 cells for each. g Representative immunoblot analysis of whole cell extracts for the indicated proteins from HeLa cells treated with siControl or siDONSON for 48 h. HSP90 was used as a loading control. h Representative immunoblot analysis of whole cell extracts for ETAA1 upon treatment with a proteasome inhibitor in DONSON-depleted cells. HeLa cells were treated with siControl or siDONSON for 48 h and 20 μM MG132 for 6 h. α-Tubulin was used as a loading control. i Ubiquitination of ETAA1 upon DONSON depletion. HeLa-Tet3G mNG-ETAA1 cells were treated with siControl or siDONSON and doxycycline (Dox) for 48 h, transfected with a HA-ubiquitin expressing plasmid for 24 h, and treated with 20 μM MG132 for 6 h. mNG-ETAA1 was immunoprecipitated with mNG-beads. Immunoprecipitates were blotted with HA antibody to detect ubiquitylated ETAA1. IP, immunoprecipitation; WCE, whole cell extract. j, l Inducible expression of ETAA1 in DONSON-depleted cells. HeLa-Tet3G mNG-ETAA1 cells were treated with siControl or siDONSON following Dox treatment for 24 h. Arrowheads indicate centrosomes. k, m Histograms represent the frequency of S-phase cells with the indicated phenotypes observed in (j) and (l), respectively. n  =  3 independent experiments, 30 cells for each. All scale bars, 5 μm. Values are mean percentages ± s.d. Tukey’s multiple comparison test was used in (f), and a two-tailed, unpaired Student’s t-test was used in (k) and (m) to obtain P-value. Source data are provided as a Source Data file.

Fig. 5. DONSON stabilizes the CMG helicase during S phase to maintain the intrinsic S/G2 checkpoint.

a Protein expression levels of ETAA1 in G1, S, G2, and M phases of WT cells. HeLa cells were synchronized in G2 phase with 10 µM RO-3306 (Cdk1i), released, and harvested at the indicated time points. b A schematic of the DNA replication termination based on previous studies. c Representative immunoblot analysis of whole cell extracts for ETAA1 from HeLa cells treated with 20 μM P22077 (USP7i) or 10 µM NMS-873 (p97i) for 6 h. α-Tubulin was used as a loading control. d Representative immunoblot analysis of whole cell extracts for ETAA1 from HeLa cells treated with siControl or siDONSON for 48 h, followed by treatment with DMSO or 10 µM NMS-873 (p97i) for 6 h. e The amount of chromatin-bound Cdc45 in S phase upon DONSON depletion. HeLa cells were treated with siControl or siDONSON for 48 h, and soluble proteins were pre-extracted by PBS containing 0.2% Triton X-100 on ice for 10 s before fixation. f Quantification of the normalized signal intensity of chromatin-bound Cdc45 in S phase observed in (e). n = 50 from three independent experiments. g The amount of chromatin-bound Cdc45 in S phase upon DONSON depletion. HeLa cells were treated with siControl or siDONSON for 48 h and synchronized at the G1/S boundary with 6 µM Aphidicolin for 24 h, released and harvested at the indicated time points. Soluble proteins were pre-extracted by PBS containing 0.2% Triton X-100 on ice for 10 s before fixation. h Quantification of the normalized signal intensity of Cdc45 in the chromatin observed in (g). n = 50 from three independent experiments. i The amount of chromatin-bound p97 in S phase upon DONSON depletion. HeLa cells were treated with siControl or siDONSON for 48 h, and soluble proteins were pre-extracted by CSK buffer on ice for 20 s before fixation. j Quantification of the normalized signal intensity of chromatin-bound p97 in S phase observed in (i). n = 50 from three independent experiments. k The amount of chromatin-bound Cdc45 in S phase upon DONSON depletion. HeLa cells were treated with siDONSON for 48 h, and with or without NMS-873 for 6 h, and soluble proteins were pre-extracted by PBS containing 0.2% Triton X-100 on ice for 10 s before fixation. l Quantification of the normalized signal intensity of Cdc45 in the chromatin in S phase observed in (k). n = 50 from three independent experiments. m Phosphorylation of Plk1 at T210 residue in S phase. HeLa cells were treated with siDONSON for 48 h, and with or without NMS-873 for 6 h before fixation. n Quantification of the normalized signal intensity of phosphorylated Plk1 (T210) at each centriole observed in (m). n = 50 from three independent experiments. o Schematic model of the phenotypes of DONSON depletion. During DNA replication, ETAA1 localized to chromatin through its interaction with the single-stranded DNA-binding protein RPA. Depletion of DONSON causes precocious CMG disassembly and following ETAA1 degradation in S phase. All scale bars, 5 μm. Values are mean percentages ± s.d. Mann–Whitney U test was used in (f), (j), (l), and (n), and the Kruskal–Wallis test was used in (h) to obtain the P-value. Source data are provided as a Source Data file.

Fig. 6. DONSON stabilizes the CMG helicase by suppressing precocious loading of the Cul2-LRR1 complex onto the CMG helicase.

a The Predicted Aligned Error (PAE) plot of the AlphaFold-Multimer (AF-M)-predicted DONSON-MCM3 and DONSON-MCM6 complex structure. b Ribbon representation of the predicted DONSON-MCM3 complex structure corresponding to (a). c HEK293T cells co-expressing MCM3-3×FLAG and HA-DONSON or the indicated deletion mutant were immunoprecipitated with FLAG antibodies. d A schematic of MCM7 poly-ubiquitination mediated by Cul2-LRR1 E3-ubiquitin ligase based on the previous studies. e The Predicted Aligned Error (PAE) plot of the AlphaFold-Multimer (AF-M)-predicted complex structure of MCM3/5/7-LRR1-ELOB-ELOC-Cul2-RBX1 with or without DONSON. f Representative structural models of the predicted protein complexes corresponding to the PAE plots in (e). All proteins are indicated in different colors. In the protein complex on the right, which lacks DONSON, the primary ubiquitination sites of MCM7, K26 and K27 (light blue), can be in close proximity (approximately 50 Å) to the C-terminal region of Cul2 and RBX1. g The interaction between endogenous MCM3 and LRR1 upon DONSON depletion. HeLa cells were treated with siControl or siDONSON for 48 h, and endogenous MCM3 was immunoprecipitated with an antibody against MCM3. h The amount of chromatin-bound LRR1 in S phase upon DONSON depletion. HeLa cells were treated with siControl or siDONSON for 48 h, and soluble proteins were pre-extracted by CSK buffer on ice for 10 s before fixation. i Quantification of the normalized signal intensity of chromatin-bound LRR1 in S phase observed in (h). n = 50 from three independent experiments. j Inducible expression of RNAi-resistant mNG-DONSON or the indicated deletion mutant upon DONSON depletion. k Quantification of the normalized signal intensity of phosphorylated Plk1 (T210) at each centriole with the indicated treatment. n = 50 from three independent experiments. All scale bars, 5 μm. Values are mean percentages ± s.d. Mann–Whitney U test was used in (i) and the Kruskal–Wallis test was used in (k) to obtain the P-value. Source data are provided as a Source Data file.

Fig. 7. DONSON maintains proper cell division by coordinating DNA and centrosome replication cycles.

a Time-lapse observation of mitotic cell division in DONSON-depleted cells. HeLa GFP-centrin-1 cells were visualized every 20 min for 48 h after 24 h treatment with siRNA and 3 h treatment with 100 nM SiR-DNA. Left–right arrows indicate precociously disengaged centrioles. b Histograms represent the frequency of cells in mitosis with the indicated phenotypes observed in (a). n  =  3 independent experiments, 30 cells for each. c Mitotic cell division of hTERT-immortalized fibroblasts derived from the patients with mutations in DONSON. Fibroblasts were infected with pMSCV-empty-vector or pMSCV-DONSON and immunostained with antibodies against Cep192 (red). d, f Inducible expression of ETAA1 upon DONSON depletion. HeLa-Tet3G mNG-ETAA1 cells were treated with siControl or siDONSON following Dox treatment for 24 h. e, g Histograms represent the frequency of cells in mitosis with the indicated phenotypes observed in (d) and (f), respectively. n = 3 independent experiments, 30 cells for each. h A speculative model showing how DNA replication machinery exerts direct control over the licensing and execution of centrosome duplication. All scale bars, 5 μm. Values are mean percentages ± s.d. Two-tailed, unpaired Student’s t-test was used in (b), (e), and (g) to obtain the P-value. Source data are provided as a Source Data file.